ABSTRACT
There are a number of multivariate techniques which extend the analytic methods given in the rest of the book to cover situations in which more than one DV is included or to other more complex data sets. They are more complicated to conduct and to interpret than the other techniques, and they involve more decisions about how the data will be treated. Such decisions can be made either by the researcher or by a computer program. They can be subject to inappropriate use, or they may capitalise on chance and give a solution which is only applicable to the given data and not provide a reliable model. The particular decisions made, either by researcher or computer, should be fully reported, in order that the reader may put the results in the context of those decisions. They generally require a much larger sample size, both for power and to produce a reliable analysis, than their equivalent univariate technique.
